Transcript Spatially explicit population models

Linking Metapopulations to
Landscape Ecology
New Paradigms
New Computers
New Software
GPS
Radio Telemetry
Spatially-Explict
Population Models
(Dunning et al. 1995)
Common language
Meaningful Interface
between Conservation
Biology, Population Ecology,
and Resource Management
Topics
• Basics of Modeling
• Sensitivity Analysis
• Use in Wildlife Management
–
–
–
–
Predicting response to habitat change
Determining harvest
Anticipating exotic species spread
Conserving and reintroducing endangered
species
– Assessing risk
SEPMs
• Goal: Model population dynamics in
realistic setting so that effects of various
land management options or global change
scenarios on wildlife populations can be
appraised
• Data Intensive: Habitat-specific
demography, dispersal, habitat selection
Two Approaches
• Each cell contains an individual (cell size =
territory or home range size)
– Equations to describe daily decisions to forage, avoid
predators, etc that eventually is translated to individual
growth
– Equations to describe yearly decisions to breed or
disperse and tie this to survival and reproductive output,
then to lambda at annual time intervals
• Each cell contains a population (cell size is a
deme)
– Equations to model birth, death, immigration,
emigration
Sensitivity Analysis
• Often most important output of model when
knowledge is less than perfect
– Determine relative importance of model parameters to
population change
• Focus future research
• Identify processes of management concern
• Bachman’s Sparrows (Pulliam et al. 1992)
–
–
–
–
–
Adult Survival
Juvenile Survival
Mature Habitat
Reproduction
Dispersal Mortality
Sx = 1.4
Sx = 1.2
Sx = 0.92
Sx = 0.51
Sx = 0.04
Timber
Harvest and
Bachman’s
Sparrow (Liu
et al. 1995)
Non-target management
species (target was
RCWP)
Simulating Effects of
Various Harvest Scenarios
Through Time
Sparrows will return to
acceptable level after going
through a bottleneck, but
return time depends on timber
harvest scenario—random
harvest is poorest scenario
Historic Change In Moose
Density (McKenney et al. 1998)
Identifying where
populations have increased
or decreased through time
helps managers decide
where to harvest in future
Predicting Change in Spread of
Exotic Species (Rushton et al. 1997)
Gray Squirrel
(Pest in UK)
Red Squirrel
(Native in UK)
Determining
Habitat
Needs of an
Endangered
Species
(Letcher et al. 1998)
Model of extremely
complex sociality
of Red-cockaded
Woodpeckers
Sensitivity Analysis
Fledgling Production
Female Breeder
Mortality
Female Disperser
Mortality
have strong effects on
population growth
Spatial Arrangement of Habitat is
Likely Important
Population growth
depends on number
of territories (amount
of habitat) and how
they are arranged
(clumping of habitat)
Research at Larger Scale Shows Importance of Dispersal
Habitat to Bighorn Management
Extinct
Extant
•Intermountain travel
corridors needed
•Domestic sheep
free to decrease
disease spread
•Focus traditionally at
the local scale
•need to switch to
metapopulation
scale
•Sheep in Santa Catalina Mountains (Arizona) are likely to be next to go extinct
•management of herd and local habitat not enough
•settlement, roads, agriculture have made intermountain travel nearly
impossible (Krausman 1997)
Ecological Niche Models
• Model species’ ecological niches
and predict geographic
distribution (Martinez-Meyer et
al. 2006)
– Ecological conditions that can
maintain populations without
immigration
• Model the conditions where species
occurs and use these to predict
where similar conditions are met
elsewhere
• Assess how these conditions change
with extrinsic factors like climate,
landcover conversion, etc.
Fig. 1 Known occurrence points (circles) of California
condor in southern California, and results of GARP
analysis predicting the potential geographic distribution
south to northern Baja California. Confidence in prediction
of potential presence is shown as a greyscale
gradient from white (no confidence) to black (high
confidence). Inset shows final areas (dark polygons)
selected as optimal for reintroductions: areas predicted
habitable at present and not in the future are shown in light
grey; areas predicted habitable at present and in the future
are in black; areas predicted not habitable at present but
that are predicted to become habitable in the future are in
dark grey.
Example of finding Reintroduction
Sites for Wolves and Condors
Fig. 2 Process of identifying suitable areas
for reintroductions of California condors
in Mexico: (A) raw GARP prediction that
reflects overall suitability of climates and
landscapes, (B) cutting by distribution of
primary vegetation in the region, (C)
weighting by distance to human presence
(roads and settlements), and (D) weighting
by future climate suitability. Confidence in
prediction of potential presence is shown
as a greyscale gradient from white (no
confidence) to black (high confidence).
Choice and Risk Models of Habitat
Quality
• Nielsen et al. article on grizzly conservation
in Alberta (required read and for discussion)
• Risk to nesting marbled murrelets
Days to predation for all
eggs (depredated and nondepredated eggs)
(the darker the color the lower the
predation)
Days to pred = 8.04 – 8.16 landscape
patch density at 5km + 1.10
landscape contrast weighted edge
density at 2km – 10.31 ShannonWeaver evenness index at 2km (r2 =
0.27)
Stands used to test model
(2000 and 2001)
Test of models observed vs. predicted:
1. Percent eggs: r= 0.22, p=0.19
2. Percent eggs and chicks
depredated: not tested because
only eggs used to test model
3. Days to predation: r=0.29, p=0.07
Applicability of model:
Songbird study sites
(green low, blue high
predation)
•Low n = 8
•High n = 9
Species Studied
Sub-Canopy:
• Pacific-slope Flycatcher
•American Robin
Shrub:
• Swainson’s Thrush
Ground:
•Wilson’s Warbler
•Song Sparrow
• Dark-eyed Junco
• Winter Wren
Surveys
– Point count predators
(Luginbuhl et al. 2001)
– Point counts for songbirds
– Map location of all
detections every 2 weeks
– Spot map songbird breeding
behavior (Vickery Index of
Success: cumulative over
season; rank 1 – 7)
– Find and monitor nests
(Mayfield mortality rates)
High
PREDATOR SPECIES
All Predators
18
Towsends's
Chipmunk
Steller's Jay
Gray Jay
Douglas
Squirrel
Low
Common
Raven
16
American
Crow
MEAN OBSERVATIONS
Predators more numerous in high predation
areas (n = 8)
p = 0.07
14
12
10
8
6
4
2
0
Nesting success differs for the American robin
and subcanopy nesters
Low
80
p = 0.003
11
31
60
4
16
10
12
p = 0.01
58
33
17
40
4
20
6
4
SPECIES
Low
Canopy
Swainson's
thrush
Song
sparrow
Pacificslope
flycatcher
Dark-eyed
junco
0
American
robin
% DEPREDATED NESTS
High
Discussion
• In small groups consider the Nielsen et al.
paper
– How did they model habitat quality?
– How does habitat quality influence their
proposed conservation strategy?
– Is the conservation strategy realistic to apply in
Alberta?
– What would you do next to improve grizzly
models?
References
• Rushton, SP, Lurz, PWW, Fuller, R., and PJ Garson. 1997. Modelling
the distribution of the red and grey squirrel at the landscape scale: a
combined GIS and population dynamics approach. J. Animal Ecology
34:1137-1154.
• Liu, J., JB Dunning, Jr., and HR Pulliam. 1995. Potential effects of a
forest management plan on Bachman’s Sparrows (Aimophila
aestivalis): Linking a spatially explicit model with GIS. Conservation
Biology 9:62-75.
• Pulliam, HR, JB Dunning, Jr., and J. Liu. 1992. Population dynamics
in complex landscapes: a case study. Ecological Applications 2:165177.
• Dunning, JB, Jr., DJ Stewart, BJ Danielson, BR Noon, TL Root, RH
Lamberson, and EE Stevens. 1995. Spatially explicit population
models: current forms and future uses. Ecological Applications 5:3-11.
• McKenney, DW, Rempel, TRS, Venier, LA, Wang, Y, and AR Bisset.
1998. Development and application of a spatially explicit moose
population model. Canadian Journal of Zoology 76:1922-1931.
• Martinez-Meyer, E., Peterson, A. T., Servin, J. I., and L. F. Kiff. 2006.
Ecological niche modelling and prioritizing areas for species
reintroductions. Oryx 40:411-418.